A Business Model Research Schema

Transcription

1 A Business Model Research Schema Susan Lambert February 2006 SCHOOL OF COMMERCE RESEARCH PAPER SERIES: 06-2 ISSN: Abstract This paper suggests a schema for business model research that has the potential to progress the research, in a structured manner, from conceptual to theoretical. It draws on the scientific and business research literature to identify the types of research necessary to further knowledge and promotes the combined use of both inductive and deductive research methods. The importance of conducting empirical research to evaluate current conceptualisations of business models and developing a theory of business models is stressed. An important aspect of any research agenda is the creation of a general classification of domain objects that can serve a wide range of current and future uses. Classification literature relating to the biological, behavioural, organisational and social sciences has been referenced in this paper in support of this claim. Existing classifications of business models are evaluated, determining that the only classifications that have been proposed to date are typologies and that no general taxonomy of business models currently exists. 1

2 Current Business Model Research As with all scientific research the overall objective of business model research is to develop a theory of business models that can aid researchers and practitioners in explaining phenomena and making predictions. Theory development is an ongoing process that occurs over long periods of time, however business model research is in its very early days. The term itself only emerged during the end of the twentieth century and the focus has been on defining business models and identifying the elements of business models (Pateli 2002; Pateli and Giaglis 2003; Osterwalder, Pigneur et al. 2005). What has emerged is an array of conceptualisations of business models that have been conceived from differing views of the problem domain. Some definitions are quite abstract and outward looking (Timmers 1998; Hamel 2000; Rappa 2001; Weill and Vitale 2001; Hawkins 2002) whilst others are quite detailed and all encompassing of business functions (Chesbrough and Rosenbloom 2000; Mahadevan 2000; Dubosson-Torbay, Osterwalder et al. 2002; Osterwalder, Pigneur et al. 2005). One of the explanations for this is that the definitions and the attributes of the business model are established according to the business model s intended use. Some researchers require the business model to form the basis of enterprise models and therefore include the internal workings of the organisation whereas others require the business model to focus on the relationships with external entities within their domain (Lambert 2003). Osterwalder, Pigneur et al. (2005), propose a business model ontology that consists of nine business model elements derived from a consolidation and refining of earlier researcher attempts at identifying business model elements. (See Osterwalder 2004 for a detailed discussion of the process.) It has been conceived with a view to creating concepts and tools that help [the] manager to capture, understand, communicate, design, analyze, and change the business logic of their firm (Osterwalder, Pigneur et al p.19). Unlike the earlier attempts at business model conceptions, this ontology draws extensively and systematically on the foregoing research. 2

3 Overview of the Research Problem and the Business Model Research Schema (BMRS) Business model research is stalled at trying to decide what attributes should be included in a business model. To date the research has been almost exclusively of a conceptual nature and this has created a problem of there being a myriad of concepts, ontologies and frameworks of business models all of which have merit, but none of which have been universally accepted. Hanks, Watson et al. (1993 p.11) recognised a similar problem in relation to research into organisational life cycle models. They concluded that due to the absence of careful empirical analysis, a plethora of conceptually based models have emerged, and that evaluation of the models can only be achieved through systematic empirical research. Small enterprise and accounting researchers found themselves in a similar position a decade or so ago and recognised that progress would only be made through exploratory, inductive, empirical research (McMahon 1998). In this paper a distinction is made betweeen concepts and theories. Concepts are the building blocks of theory. A concept is an idea expressed as a symbol or in words (Neuman 2003 p.44). [Concepts] are used to communicate the essence of an observation (Cavana, Delahaye et al p33). A theory is a set of systematically interrelated concepts, definitions, and propositions that are advanced to explain and predict phenomena (facts) (Cooper and Schindler 2000 p.51). The objective of this paper is to propose a schema for research into business models that has the potential to progress future research towards a theory that can assist in explaining and predicting phenomena in relation to business models. An important aspect of that schema is the creation of a general taxonomy of business models. In the following 3

4 section the relevance of scientific research methods for business model research is discussed along with the role that classification plays in theorising. The differences between the specific and general classification schemes and between typologies and taxonomies are then explored along with an analysis of current business model classifications. It is important to keep moving the research from conceptual frameworks to theoretical frameworks so that explanatory and predictive research can be undertaken. For this to occur an understanding of business model variables and their relationships with each other is required. The source of this understanding lies with the gathering and analysis of empirical data, specifically through the construction of a general taxonomy. Figure 1 illustrates a research schema that has been designed to incorporate both deductive and inductive research methods and depicts the overall goal of developing theories about business models. Figure 1 depicts the six phases of business model research beginning with simple conceptualisations about business models and culminating in the development of theories of business models. Each phase is a necessary step in the progression of research and contributes something to the following steps. It is important point to note that the steps are iterative, empirical findings cause a return to conceptualisations and inductive research feeds into future deductive research. Phase 1: Early Conceptualisation of Business Models Existing research into business models falls predominantly in this phase of research. The research consists of simple conceptualisations of business models including definitions, the identification of business model elements and typologies of business models. Phase 2: Deductive Empirical Research Very little empirical research has been conducted into business models. There have been small amounts of case study research and some application of typologies all of which represent deductive research. The conceptualisations have been used as the basis for classifying the data. 4

5 Phase 1 Early conceptualizations of business models Phase 2 Deductive empirical research Phase 3 Developed conceptualizations of business models Phase 4 Inductive empirical research Phase 5 Generalisations Phase 6 Theories of business models Figure 1: Business Model Research Schema Phase 3: Developed Conceptualisations of Business Models As a result of the deductive empirical research and new contributions to business research literature, the original concepts are developed and refined. Two types of research emerge from these developed concepts; further deductive research can be conducted (back to Phase 2) or inductive empirical research can be undertaken (Phase 4). It will be argued throughout this paper that it is necessary to undertake inductive empirical research in order to move away from conceptualisation and towards theories of business models. 5

6 Deductive research serves to improve the business model concepts but cannot provide the basis for theory development. Phase 4: Inductive Empirical Research This phase consists of taxonomic research. The selection of variables for which data will be collected needs to be all encompassing. All relevant business model variables need to be identified thereby reducing the chance of omitting important data. Selection of the variables is based on the foregoing research, in particular the business model elements identified in the research conducted at Phase 3. The data are analysed using quantitative data analysis techniques including multivariate analysis and a taxonomy is produced. The taxonomy will provide the means by which business models can be classified and named and relationships between variables can be measured. The taxonomy will generate knowledge that can be used to refine the existing conceptualisations of business models (Phase 3) and to provide the basis of the next phase, namely Generalisations. Phase 5: Generalisations General patterns of configurations of business model variables can be inferred from the taxonomy. Simple relationships between variables can be hypothesised and tested. The generalisations will feed into Phase 6 forming the basis of a theory of business models. Phase 6: Develop Business Model Theory The aim of a theory is to present a systematic view of phenomena by specifying relations among the variables with the purpose of explaining and predicting the phenomena. (Kerlinger 1974 p.11). It is important to recognise from this schema that theory development can only be achieved once sound inductive research has been conducted. The result of the inductive research is a taxonomy of business models that will provide the necessary classifications of business models on which to generate theories. The following sections explain the 6

7 roles of inductive and deductive research and the significance of the distinction between typologies and taxonomies. The Inductive-Deductive Research Cycle Research methods can be categorised in many ways but one that appears important in analyzing business model research is according to the direction of the reasoning process. Deductive reasoning begins with an abstract concept and then tests that concept with empirical evidence. Support for the concept is achieved if data collected from observations are consistent with the proposed concept. An example of deductive research into business models is Weill and Vitale s (2002) study of IT infrastructure focus in each of their atomic business model classes. In prior research, Weill and Vitale conceptualised eight different business model types which they called atomic business models. They then hypothesised that the IT requirements of each atomic business model would be different. To test this concept they conducted empirical research in the form of interviews with senior managers to ascertain the atomic business models being implemented and the IT infrastructure focus. The data were analysed, and it was deduced that each atomic business model had a different IT infrastructure focus. Inductive research on the other hand creates grounded theory by beginning with data collection and then making generalisations and inferring theories based on the observations (Cavana, Delahaye et al. 2003; Neuman 2003). Inductive research should begin with no preconceived ideas of what will be found. An inductive research question might be, What business model attributes are associated with various IT infrastructure foci? Sound scientific research requires a combination of inductive and deductive methods to advance knowledge. The Scientific Revolution, that took place around the end of the fifteenth century, linked the rational and the empirical, thought and fact, theory and practical experiment. (Bronowski 1951 p.31). What this means is that knowledge is 7

8 advanced by performing deductive research based on the results of the inductive research and that these steps require iteration. Ghent (1966) suggests that experimental design in the biological sciences begins with guesswork or a concept based on existing knowledge of the researcher. This concept forms the basis of the design of empirical experiments from which the original concept can be modified (deductive research). The modified concept (the result of deductive research) should be more realistic than the original concept and will form the basis of inductive research. This inductive-deductive cycle provides the means by which concepts can be developed into theories. To be of value, research must progress from being descriptive and conceptual to being explanatory and predictive. It must move from conceptual frameworks towards theoretical frameworks. The goal of natural science, according to March and Smith (1995 p.253), is to develop theories that provide deeper, more encompassing, and more accurate explanations [of reality]. This goal is equally relevant to the social sciences and to business model research. The BMRS incorporates both deductive and inductive research methods and reflects the iterative nature of the process and the progression towards theory development. The Need for a Business Model Classification Scheme To advance from concept to theory, it is necessary to order or classify the objects within the domain. Classifications are partway between a simple concept and a theory. They help to organise abstract, complex concepts (Neuman 2003 p.46). Business models are abstract, complex concepts of which understanding can be enhanced through the development of a general classification scheme. Classification involves the ordering of objects into groups or classes on the basis of their similarity (Bailey 1994). This ordering of objects into classes provides meaning to reality (Simpson 1961). It also aids our understanding of a domain as we do not perceive, 8

9 remember and talk about each object and event as unique, but rather an instance of a class or concept that we already know something about (Smith and Medin 1981 p.1). Just as recognition of similarities and differences among objects and classification of objects are fundamental steps in the development of childhood reasoning (Piaget 1973) the recognition of similarities and differences between business models and the development of classes of business models are fundamental to business model research. In all forms of scientific research, including social science research, classification of objects within the domain is an important step towards other research (Mezzich and Solomon 1980; McKelvey 1982). A good classification scheme forms the foundation of theory development. Theory cannot explain much if it is based on an inadequate system of classification (Bailey 1994 p.15). Bailey (1994) identifies ten advantages of a good classification scheme: 1. Provides an exhaustive and perhaps even definitive array of types or taxa. 2. Reduces complexity and achieves parsimony. 3. Identifies similarities among objects and allows a group of objects to be analysed at the exclusion of other, more diverse objects. 4. Identifies differences so that dissimilar objects can be separated for analysis. 5. Presents an exhaustive list of dimensions or characteristics. 6. Allows types of objects to be compared. 7. Manages and takes stock of types of objects. 8. Enables the specification of hypotheses concerning relationships between classes of objects through the creation of typologies and then identification of empirical cases. 9. Types can be used as criterion for measurement. One type can be used as the reference point and others can be measured relative to that criterion. 10. Provides versatility, i.e. suits many needs and can display different aspects of the data. 9

10 A good classification scheme organises objects according to their place within the problem domain and depicts relationships between objects. Gilmour is of the belief that The primary function of classification is to construct classes about which we make inductive generalisations (Gilmour 1951 p. 401). Typologies and Taxonomies Classifications may be designed to serve multiple purposes or very specific purposes. Those that serve many purposes and provide more meaningful generalisations are referred to as natural or general classifications and those that serve specific or few purposes are referred to as artificial or arbitrary classifications (Simpson 1961; Sokal and Sneath 1963). McKelvey (1982) in discussing organisational systematics, recognises two basic classification schemes; special classifications and general classifications. Special classifications focus on only one or a few attributes of interest. They are developed with a special purpose in mind and therefore have limited utility. In contrast, general classification schemes attempt to group objects based on all of their attributes. Another important distinction in classification schemes is between typologies and taxonomies although many researchers use the terms interchangeably. Understanding the differences between taxonomic research and typological research is important because it will be seen that they serve different purposes and have their own limitations and strengths. Typologies Typologies are a product of deductive research. The researcher conceptualises the types that are relevant to the research. These types form the cells of the classification scheme and each cell is labeled (named). The researcher then identifies cases that possess the characteristics deemed essential to fit the cells. 10

11 The great advantage of typologies is their ability to simplify complex concepts by classifying objects according to a few, often two, criteria at a time. Furthermore A sound typology forms a solid foundation for both theorising and empirical research (Bailey 1994 p 33). Typological research does however have its limitations. Hambrick (1984 p.28) points out that: typologies represent a theorist s attempt to make sense out of non-quantifiable observations they are largely the product of rather personal insight, they may not accurately reflect reality. Or more likely they may serve well for descriptive purposes but have limited explanatory or predictive power Typologies are specific classifications rather than general classifications. They serve limited purposes, therefore it is conceivable that over time a large number of typologies will be developed, each capable of serving a specific purpose. Each typology will add something to the business model knowledge base however it will be difficult, if not impossible, to consolidate the research findings due to their specific nature. Typologies are mostly generated through qualitative classification rather than quantitative or statistical analysis (Bailey 1994). Bailey (1994) points out however that typologies can be formed through conceptualising the types and then using cluster analysis to quantify the empirical findings. Taxonomies Taxonomies, in contrast to typologies, are generated from inductive research, derived empirically (Sokal and Sneath 1963) through multivariate analysis (Hanks, Watson et al. 1993). The researcher creates grounded theory by first collecting the data and then generalising to the abstract or conceptual. A taxonomy begins empirically, rather than conceptually, with the goal of classifying cases according to their measured similarity on observed variables. (M S Lewis-Beck, Series Editor in Bailey 1994 p.v). Unlike typologies whereby the categories are derived conceptually, taxonomic categories are derived through cluster analysis. 11

12 The term taxonomy, like classification, can be used to refer to a process and the end result. Taxonomy [the process] is the theoretical study of classification, including its bases, principles, procedures and rules (Simpson 1961 p. 11). Taxonomic research involves the identification of a large number of variables on which to gather data. Determination of these variables must be based on existing domain knowledge that has been generated through deductive research. The data are analysed using multivariate analysis to identify the natural groupings (classes). The aim is to minimise within group variance and maximise between group variance thereby creating homogeneous groups. Once these homogeneous groups are created they can be used for a multitude of research applications; within group behaviour can be studied as well as inter-group behaviour. The Role of Typologies and Taxonomies in the BMRS Both typologies and taxonomies have a role to play in the BMRS but it is important that their output is applied appropriately. Table 1 summarises the differences between typologies and taxonomies and their role in the BMRS. In terms of the BMRS, typologies are produced at Phase 2 providing input to refine conceptualisations of business models and as foundations for identifying relevant business model variables for future inductive research. The generation of a business model taxonomy through inductive research falls within Phase 4 of the BMRS. The taxonomy would provide a basis for classifying business models according to their degree of affinity; in other words, according to the extent of similarity of all their characteristics. As new instances of business model are incorporated into the database the taxonomy can be updated. This classification scheme will be selfadjusting over time. As the data is added to or changed and the analysis reperformed, the resulting classes will change. 12

13 Typologies Falls within Phase 2 on the BMRS Specific/arbitrary/artificial classification Derived conceptually Form of deductive research Reasoning by deduction Few characteristics considered Mostly qualitative classifications Provides a basis for only limited generalisations Taxonomies Falls within Phase 4 of the BMRS General/natural classification Derived empirically Form of inductive research Reasoning by inference Many characteristics considered Quantitative classifications Provides a basis for generalisation Table 1: Summary of differences between Typologies and Taxonomies An Analysis of Exisiting Business Models Research using the BMRS It was recognised by early business model researchers (Bambury 1998; Timmers 1998; Rappa 2001) that an understanding of business models requires some form of classification. The problem is that the early classifications were based on nothing more than anecdotal evidence and conceptualisations of business models (Hawkins 2002). More recently case study research has been undertaken in an attempt to support conceptualisations of business models (Osterwalder, Pigneur et al. 2005) and some typological research is being conducted (Weill and Vitale 2002; Weill, Malone et al. 2005). These classification schemes add to the body of knowledge and assist in the identification of potential variables for future data collection. 13

14 Currently there exists no taxonomy of business models (Osterwalder, Pigneur et al. 2005; Weill, Malone et al. 2005). To progress business model research it is necessary to collect large amounts of data on as many business model elements (variables) as possible and to construct a taxonomy of business models. A comprehensive analysis of business model literature has been presented by Pateli and Giaglis (2003) in what they refer to as an explanatory framework consisting of six categories of research sub-domains, namely: Definitions Components Taxonomies Representations Change Methodologies Evaluation Models In terms of the BMRS proposed in this paper the Definitions and Components research clearly fall into the Early Conceptualisations (Phase1) of the BMRS. It could be argued that the remaining sub-domains fall into Phase 3 since they are built on the earlier research, however some of this research could be classified as early conceptualisations and also form part of Phase 1. Although Pateli and Giaglis (2003) call one of their subdomains Taxonomies they state in the description of the sub-domain that it consists of various typologies of business models. The research referred to consists of conceptualisations of typologies which therefore falls into Phase 1 of the BMRS. None of the research sub-domains represent empirical research. Like Pateli and Giaglis (2003), Osterwalder et al. (2005) have provided a detailed analysis of existing business model research. Three categories of business model research are presented in a hierarchy referred to as a Business Model Concept Hierarchy. It describes the relationships between business model definitions and meta-models, the 14

15 classifications of the meta-models into business model types, and the identification of existing instances of the business model types. The definitions of business models and the creation of meta-models equate to Phases 1 and 3 of the BMRS depending on how developed the concepts are. Osterwalder et al. (2005) use the term taxonomy to refer to different meta-models, however it is in fact a typology of meta-models and therefore forms part of Phase 3 of the BMRS. The third level of Osterwalder et al. s (2005) Business Model Concept Hierarchy consists of identifying instances of business models types. This is a form of empirical research that aims to support the conceptualised meta-model types and is therefore deductive research and part of Phase 2 of the BMRS. Osterwalder et al. (2005) also provide a five phase evolution of business model research. The first three phases, being business model definitions, classifications, component identification and description, are all part of the BMRS Phase 1, early conceptualisations of business models. The fourth phase of Ostwalder et al. s (2005) evolution model represents developed conceptualisations (Phase 3 of the BMRS). Osterwalder et al. (2005) state that some empirical testing was carried out during this phase. This testing is considered to be deductive rather than inductive and is therefore equivalent to Phase 2 of the BMRS. Osterwalder et al. s fifth phase is the application of the research to IS and management applications. This phase does not form part of the BMRS since it does not represent scientific research, rather it is the application of the research. Summary The BMRS presented in this paper captures the essence of the Scientific Revolution. Essentially it suggests a combination of deductive and inductive research methods be employed to build empirically on existing concepts of business models and progressively refining those concepts to the point where theories of business models can be proposed. 15

16 To date, research into business models has been conceptual and any empirical research has been deductive. On close examination it has been established that the existing taxonomies of business models are in fact typologies and are of limited and very specific use. There exists no general taxonomy of business models largely because there is no widely agreed upon concept of a business model. However this should not act as a barrier to the process and at some point the conceptualising must give way to inductive research. The inductive research will give rise to new conceptualisations and eventually to theories of business models. The BMRS emphasises that the phases need to be iterated, each iteration providing more knowledge to improve the next. It is not expected that the first taxonomy of business models will be complete, but it will add to the existing body of knowledge and improve conceptualisations of business models through inference. A classification scheme, like a good theory, is seldom finished. It is only given interim acceptance with the understanding that further studies will tend to elaborate and refine it, or disconfirm it (McKelvey 1982 p.30). This paper suggests a structured research agenda to take the current deductive research efforts towards theories of business models by developing a general taxonomy of business models. It is only through theory development that explanatory and predictive questions be answered. 16

17 Reference List Bailey, K. D. (1994). Typologies and Taxonomies: An Introduction to Classification Techniques. Los Angeles, Sage Publications Inc. Bambury, P. (1998). "A Taxonomy of Internet Commerce." First Monday 3(10). Bronowski, J. (1951). The Common Sense of Science. London, Heineman Educational Books Ltd. Cavana, R. Y., B. L. Delahaye, et al. (2003). Applied Business Reseach: Qualitative and Quantitative Methods. Milton Queensland, John Wiley & Sons Australia. Chesbrough, H. and R. Rosenbloom (2000). The Role of the Business Model in Capturing Value Innovation: Evidence from Xerox Corporation's Technology Spinoff Companies. Boston, Massachusetts, Harvard Business School. Cooper, D. R. and P. S. Schindler (2000). Business Research Methods. New York, McGraw-Hill Irwin. Dubosson-Torbay, M., A. Osterwalder, et al. (2002). "E-business model design, classification, and measurements." Thunderbird International Business Review 44(1): Forbes, L. P., S. W. Kelley, et al. (2005). "Typologies of e-commerce retail failures and recovery strategies." The Journal of Services Marketing 19(5): 280. Ghent, A. W. (1966). "The Logic of Experimental Designin the Biological Sciences." Bioscience 16: Gilmour, G. (1951). "The development of taxonomic theory since 1851." Nature 168: Hambrick, D. C. (1984). "Taxonomic Approaches to Studying Strategy: some Conceptual and Methodological Issues." Journal of Management 10(1): Hamel, G. (2000). Leading the Revolution. Boston, Harvard BUsiness School Press. Hanks, H. S., C. J. Watson, et al. (1993). "Tightening the Life-Cycle Construct: A Taxonomic Study of Growth Stage Configurations in High-Technology Organizations." Entrepreneuship, Theory & Practice: Hawkins, R. (2002). The Phantom of the MArketplace: Searching for New E-Commerce Business Models. Euro CPR 2002, Barcelona,. 17

18 Kerlinger, F. (1974). Foundations of Behavioural Research. New York, Holt, Rienhart and Winston. Lambert, S. (2003). Making Sense of Business Models. Mahadevan, B. (2000). "Business Models for Internet-based E-commerce: An Anatomy." California Management Review 42(4): March, S. T. and G. F. Smith (1995). "Design and natural science research on information technology." Decision Support Systems 15: McKelvey, B. (1982). Organizational Systematics: Taxonomy, Evolution, Classification. Berkeley, University of California Press. McMahon, R. G. P. (1998). Business growth and performance and the financial reporting practices of Australian manufacturing SMEs. Armidale, University of New England: 444. Mezzich, J. E. and H. Solomon (1980). Taxonomy and Behavioral Science: Comparative Performance of Grouping Methods. London, Academic Press Inc (London) Ltd. Neuman, W. L. (2003). Social Research Methods: Qualitative and Quantitative Approaches. Boston, Pearson Education Inc. Osterwalder, A., Y. Pigneur, et al. (2005). "Clarifying Business Models: Origins, Present and Future of the Concept." CAIS 15. Pateli, A. (2002). A Domain Area Report on Business Models. Athens, Athens University of Economics and Business. Pateli, A. and G. Giaglis (2003). A Framework for Understanding and Analysing e- Business Models. 16th Bled Electronic Commerce Conference, Slovenia. Piaget, J. (1973). The Psychology of Intelligence. New Jersey Littlefield, Adams and Co. Rappa, M. (2001). Business Models on The Web. Managing the Digital Enterprise. North Carolina Simpson, G. G. (1961). Principles of Animal Taxonomy. New York, Columbia University Press. Smith, E. and D. Medin (1981). Categories and Concepts. Cambridge, Harvard University Press. Sokal, R. R. and P. H. A. Sneath (1963). Principles of Numerical Taxonomy, W H Freeman and Company. 18

19 Timmers, P. (1998). "Business Models for Electronic Markets." Electronic Markets 8(2): 3-8. Weill, P., T. W. Malone, et al. (2005). Do Some Business Models Perform Better than Others? A Study of the 1000 Largest US Firms. Massachusetts, Sloan School of Management Massachusetts Institute of Technology. Weill, P. and M. Vitale (2001). Place to Space. Boston, Harvard Business School Press. Weill, P. and M. Vitale (2002). "What IT Infrastructure Capabilities are Needed to Implement E-Business Models?" MIS Quarterly Executive 1(1): Wong, H. Y. and B. Merrilees (2005). "A brand orientation typology for SMEs: a case research approach." The Journal of Product and Brand Management 14(2/3):